NUMBER AUGUST 2002 CONTRIBUTIONS IN SCIENCE SYSTEMATICS OF XANTUSIID LIZARDS OF THE GENUS LEPIDOPHYMA ROBERT L. BEZY AND JOSÉ L. CAMARILLO R.

Transcription

1 NUMBER AUGUST 22 CONTRIBUTIONS IN SCIENCE SYSTEMATICS OF XANTUSIID LIZARDS OF THE GENUS LEPIDOPHYMA ROBERT L. BEZY AND JOSÉ L. CAMARILLO R.

2 SERIAL PUBLICATIONS OF THE NATURAL HISTORY MUSEUM OF LOS ANGELES COUNTY The scientific publications of the Natural History Museum of Los Angeles County have been issued at irregular intervals in three major series; the issues in each series are numbered individually, and numbers run consecutively, regardless of the subject matter. Contributions in Science, a miscellaneous series of technical papers describing original research in the life and earth sciences. Science Bulletin, a miscellaneous series of monographs describing original research in the life and earth sciences. This series was discontinued in 978 with the issue of Numbers 29 and 3; monographs are now published by the Museum in Contributions in Science. Science Series, long articles and collections of papers on natural history topics. Copies of the publications in these series are sold through the Museum Book Shop. A catalog is available on request. The Museum also publishes Technical Reports, a miscellaneous series containing information relative to scholarly inquiry and collections but not reporting the results of original research. Issue is authorized by the Museum s Scientific Publications Committee; however, manuscripts do not receive anonymous peer review. Individual Technical Reports may be obtained from the relevant Section of the Museum. SCIENTIFIC PUBLICATIONS COMMITTEE John Heyning, Deputy Director for Research and Collections John M. Harris, Committee Chairman Brian V. Brown Gordon Hendler Inés Horovitz Joel W. Martin K. Victoria Brown, Managing Editor NATURAL HISTORY MUSEUM OF LOS ANGELES COUNTY 9 EXPOSITION BOULEVARD LOS ANGELES, CALIFORNIA 97 Printed at Allen Press, Inc., Lawrence, Kansas ISSN

3 SYSTEMATICS OF XANTUSIID LIZARDS OF THE GENUS LEPIDOPHYMA ROBERT L. BEZY AND JOSÉ L. CAMARILLO R. 2 ABSTRACT. Multivariate analyses of variation in 3 scale characters among 29 locality samples of members of the genus Lepidophyma (A. Duméril in Duméril and Duméril, 85) identify the presence of 6 morphological groups in México below the latitude of 9 N. Univariate comparisons among the samples diagnose a total of 7 morphological species in the genus, of which occur in southern México. Three species of Lepidophyma have broad (but discontinuous) geographic distributions: L. flavimaculatum A. Duméril in Duméril and Duméril, 85 (Veracruz to Panamá), L. smithii Bocourt, 876 (Guerrero to El Salvador), and L. sylvaticum Taylor, 939 (Nuevo León to Veracruz). The other 4 members of the genus seem to have relatively small ranges, and 7 of these are known from only one or two localities. The greatest species diversity occurs in Oaxaca and Chiapas ( species) and the northern Sierra Madre Oriental of Querétaro and San Luis Potosí (4 species). RESUMEN. Un análisis multivariado de variación en 3 caracteres de escamas de 29 localidades con muestras de miembros del género de Lepidophyma (A. Duméril in Duméril y Duméril, 85) identifica la presencia de 6 grupos morfológicos en México por debajo de la latitud de 9 N. Comparaciones univariadas entre todas las muestras diagnostican un total de 7 especies morfológicas en el género, de las cuales se encuentran en el sureste de México. Tres especies de Lepidophyma tienen una amplia (pero discontinua) distribución geográfica: L. flavimaculatum A. Duméril in Duméril y Duméril, 85 (Veracruz a Panamá), L. smithii Bocourt, 876 (Guerrero a El Salvador) y L. sylvaticum Taylor, 939 (Nuevo León a Veracruz). Los otros 4 miembros del género paracen tener áreas de distribución relativamente pequeñas y7deéstas se conocen unicamente de una o dos localidades. Las regiones con máxima diversidad de especies están en Oaxaca y Chiapas ( especies) y el norte de la Sierra Madre Oriental en Querétaro y San Luis Potosí (4 especies). INTRODUCTION The living members of the family Xantusiidae range from the southwestern U.S. to Panamá and Cuba and comprise three genera: Xantusia Baird, 859, Lepidophyma A. Duméril in Duméril and Duméril, 85, and Cricosaura Gundlach and Peters in Peters, 863. The majority of the diversity exists in the Middle American genus Lepidophyma, with 5 to 2 nominal species scattered from Nuevo León, México, to Panamá. Like other xantusiid lizards, members of the genus Lepidophyma are secretive or reclusive and are seldom observed outside rock crevices, caves, and decaying tree stumps and logs. Most species are uncommon and/or difficult to collect. Their geographic distribution is spotty, with known localities often separated by considerable distances and represented by few museum specimens. The distributional gaps may be the result of several factors, including narrow habitat requirements, rugged topography, forest destruction, and inadequate field sampling. That many of the disjunctions may be real rather. Herpetology, Natural History Museum of Los Angeles County, Los Angeles, California, Laboratorio y Colección de Herpetología, CyMA, Escuela Nacional de Estudios Profesionales Iztacala, A.P. 34, Tlalnepantla, Estado de México, México. Deceased 22. than artifactual is suggested by high incidence of morphological divergence among the localities. Since the description of Lepidophyma (Duméril and Duméril, 85), a total of 25 taxa have been named in the genus, most of them based on one or two specimens from a single locality. Because many of the known localities are disjunct, represented by small sample sizes, and have various degrees of morphological differentiation, it remains unclear how many valid species exist in the genus, particularly in México (e.g., Flores-Villela, 993; Liner, 994; Smith, 973). In previous studies, we have delineated species of Lepidophyma in northeastern México (Bezy, 984), Central America (Bezy, 989), and within the cluster of species containing L. gaigeae Mosauer, 936 (Bezy and Camarillo, 992, 997). In this paper we focus on delineating the morphologically diagnosable species found in México below the latitude of 9 N and synthesize the scattered information on the genus by providing summaries, diagnoses, and a key for all species of Lepidophyma. MATERIALS AND METHODS Data were taken for 3 scalation characters from a total of,42 specimens (including all holotypes) from 6 localities throughout the range of Lepidophyma (see Specimens Examined). As in previous studies (Bezy, 984, 989; Bezy and Camarillo, 992, 997), we used discrim- Contributions in Science, Number 493, pp. 4 Natural History Museum of Los Angeles County, 22

4 2 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma inant function analysis as an exploratory tool to visualize overall patterns of variation in these characters among localities. For these analyses, the individual locality samples constituted the a priori groups, but for clarity of presentation in the resulting plots, envelopes were constructed to enclose all individuals from larger geographic areas (usually countries of Central America and states of México). After delineation of the multivariate groups, the observed limits of variation of individual univariate characters were examined to identify discrete (nonoverlapping) differences among the locality samples and sample clusters. Thus, the species are diagnosed by univariate features and are not dependent on multivariate analyses. All analyses were performed with Systat 7. (SSPS, Chicago, IL, 66 USA). It is recognized that ratios violate the assumptions of a normal distribution and can influence the results of multivariate analyses, particularly principal components analyses (e.g., Atchley et al., 976). In analyses of data, we found the results of discriminant analyses that used ratios generally to be congruent with those from analyses based on the residuals from regression of the morphometric variables on snout vent length (SVL). Despite their statistical drawbacks, we employ ratios because we found them to be more useful in recognizing and diagnosing morphospecies. We embrace the evolutionary species concept (e.g., Wiley, 978) and its operational equivalent, the phylogenetic species (e.g., Cracraft, 983). We thus recognize as separate species any locality sample or group of samples that has or more discrete difference(s) from each of the other samples or groups (i.e., for each pairwise comparison there are or more characters for which the observed limits of variation do not overlap). In most cases these morpho-species are composed of disjunct populations and probably are not united by gene flow. Whether they actually represent monophyletic groups of isolated populations is difficult to determine with the quantitative scale features utilized in this study, but the question is being addressed by phylogenetic analyses of mitochondrial DNA sequences combined with morphology (R.L. Bezy and collaborators, unpublished). In our analyses of variation, we make three exceptions to the requirement that species of Lepidophyma have discrete (nonoverlapping) differences in morphology. Two of the exceptions involve species pairs that are % diagnosable by univariate differences in areas where they occur in sympatry, but that have a small overlap in the observed limits of variation for geographically distant localities (where a combination of 2 univariate characters serves to distinguish the species). This is the operational equivalent of considering that biological species are a subset of evolutionary species. The third exception involves 2 allopatric species which have slight overlap in 2 morphological characters but have discrete differences in karyotypes throughout their ranges. We have chosen not to recognize subspecies in the genus because of the large number of geographically isolated samples with morphological features that are statistically different but that overlap the observed limits of variation in other samples. STANDARD CHARACTERS The standard data taken on all specimens consist of 3 scale characters: 2 meristic, 9 mensural, and presence/ absence. We employ the scale terminology of Savage (963) rather than that of Smith (973) in that we consider the parietals to be the scales on each side of the Figure Dorsal and lateral views of the head of a generalized specimen of Lepidophyma. Numbered points were used in scale measurements and counts (see Materials and Methods). interparietal rather than the pair posterior to it. In the descriptions below, at midbody indicates that the count was made midway between the gular fold and the vent, above a point corresponding to half the total number of transverse rows of ventrals. Pretympanics (PTMP). The minimum number of small scales (pretympanics) separating the postoculars (Fig. : 26, 27) from the second postorbital supralabial (sum of right plus left side of head). The second postorbital supralabial (Fig. : 28) is usually the seventh and is the second labial scale posterior to the eye that does not enter the orbital ring of scales. Supralabial scales are skirted and not included in the count. For example, there are a minimum of 4 pretympanics in Fig., between points 27 and 29. Gulars (GUL). The number of small scales (gulars) along the ventral midline between the posterior gular fold and the second pair of infralabials (Fig. 2, between points A and B). Gulars Contacting First Infralabials (GCIL). The number of gulars in contact medially with first pair of infralabials. (Fig. 2, gular at point C). Dorsals (DOR). The number of scales along the middorsal line between the posterior edge of the postparietals and a point directly above the vent. Dorsals Between Paravertebral Rows (DBPVR). The minimum number of dorsal scales separating the rows of large paravertebral tubercles (at midbody; Fig. 2, 5 dorsals separate tubercles D and F) Dorsals Along Paravertebral Row (DAPVR). The num-

5 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma 3 Figure 2 Ventral view of head and dorsal view of body (axilla to groin) of a generalized specimen of Lepidophyma. Lettered points were used in scale counts (see Materials and Methods). Figure 3 Dorsal (left) and ventral (right) views of the preanal area and the proximal portion of the tail of a generalized specimen of Lepidophyma with characters indicated by letters (see Materials and Methods). ber of dorsals (from above the axilla to above the groin) in the row immediately above (i.e., medial to) the paravertebral tubercle row (Fig. 2, G to H). Dorsals Between Paravertebral Tubercles (DBPVT). The number of middorsal scales equivalent to the distance between 2 consecutive large tubercles (largest size class) within the paravertebral row (at midbody; Fig. 2, the distance between tubercles E and F is equal to 4 dorsals). Large Paravertebral Tubercles (PVTL). The number of large tubercles (largest size class) in the paravertebral row (from above the axilla to above the groin; Fig. 2, between I and J there are 8 large tubercles). Paravertebral Row (PVR). The total number of scales (all sizes) in the paravertebral row (from above the axilla to above the groin; Fig. 2, there are 54 scales between I and J). Paravertebrals, Small (PVS). The number of scales (from above the axilla to above the groin) in the paravertebral row that are shorter than.5 dorsal scales. Paravertebrals,.5 (PVT.5). The number of scales (from above the axilla to above the groin) in the paravertebral row that are equal to or longer than.5 middorsal scales. Paravertebrals, 2 (PVT2). The number of scales (from above the axilla to above the groin) in the paravertebral row that are equal to or longer than 2 middorsal scales. Paravertebrals, 3 (PVT3). The number of scales (from above the axilla to above the groin) in the paravertebral row that are equal to or longer than 3 middorsal scales. Lateral Tubercle Rows (LTR). The number of large tubercular scales (largest size class) on the side of the body between the axilla and groin. The count is made along a longitudinal line that is parallel to the paravertebral row and that passes through the second large lateral tubercle below the paravertebrals (Fig. 2, there are 6 tubercle rows along line K to L). Dorsal Interwhorls (IWD2). The maximum number of annuli of small scales separating the second from the third whorl of enlarged caudal scales (at middorsum; Fig. 3, three dorsal interwhorls between M and N). Ventral Interwhorls (IWV2). The minimum number of ventrally complete annuli of small scales separating the second from the third whorl of enlarged caudal scales (Fig. 3, 3 complete ventral interwhorls between O and P). Ventrals, Longitudinal (VL). The number of transverse rows of large ventral scales between gulars and the vent, including the preanals. Ventrals, Across (VA). The number of longitudinal rows of ventrals across midbody. Longitudinal rows were excluded if (at midbody) they contain 2 more scales that are in contact with a single scale of the medially adjacent longitudinal row of ventrals. Femoral Pores (FPT). The total number of pores (right plus left femoral surface). In females and juveniles some or all of the pores may be vestigial or undeveloped (marked by shallow depressions rather than discrete glandular openings). On all specimens the counts include both developed and undeveloped pores. Fourth Toe Lamellae (FTL). The number of scales along the ventral midline of the fourth toe between the base and the claw. Divided Fourth Toe Lamellae (FTLD). The number of scales under the fourth toe with midventral sutures. Parietal Spot (PS). The presence () or absence () of a pale spot on the interparietal scale. The spot marks the approximate position of the parietal foramen (which may be roofed over by bone in adults) and is located along the dorsal midline, usually in the posterior half of the scale (Fig., point ). Prefrontal Length (RPFL). Ratio of the length along the midline of the lateral prefrontal (Fig., points 4 to 6) divided by the length along the lateral border of the lateral prefrontal (Fig. : 3 to 7). Median Prefrontal Length (RML). Ratio of the length of the median prefrontal (Fig. : 5 to 8) divided by the length of the midline suture of frontals (Fig. : 8 to 9). Specimens lacking a median prefrontal were scored as.

6 4 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma Table. Variation in 3 scale characters for 7 species of Lepidophyma. In each cell, the upper number is the mean, the middle is the standard error, and the lower are the observed limits of variation. Sample size (N) is given below the species name. Species PTMP GCIL GUL DOR DBPVR DAPVR DBPVT PVTL PVR PVS PVT.5 PVT2 PVT3 LTR IWD2 IWV2 L. dontomasi 29 L. gaigeae 77 L. lowei 6 L. radula L. smithii 397 L. tarascae 4 L. lineri 3 L. mayae 6 L. pajapanense 22 L. tuxtlae 97 L. chicoasense 4 L. lipetzi

7 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma 5 Table. Continued. Species PTMP GCIL GUL DOR DBPVR DAPVR DBPVT PVTL PVR PVS PVT.5 PVT2 PVT3 LTR IWD2 IWV2 L. flavimaculatum 433 L. reticulatum 65 L. micropholis 3 L. occulor L. sylvaticum

8 6 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma Table. Extended. Species VL VA FPT FTL FTLD PS RPFL RML RMW2 RNL RAPPSL RPNH RPOL RSLH RPAW L. dontomasi 29 L. gaigeae L. lowei 6 L. radula L. smithii 397 L. tarascae L. lineri 3 L. mayae L. pajapanense 22 L. tuxtlae L. chicoasense 4 L. lipetzi

9 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma 7 Table. Continued Extended. Species VL VA FPT FTL FTLD PS RPFL RML RMW2 RNL RAPPSL RPNH RPOL RSLH RPAW L. flavimaculatum 433 L. reticulatum 65 L. micropholis 3 L. occulor L. sylvaticum

10 8 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma Table 2. Variation in scale characters among samples of group S of Lepidophyma. The observed limits of variation are followed by the mean (in parentheses). N FPT LTR GUL DOR L. tarascae Michoacán (6.8) 6 25 (8.8) 4 43 (43.8) (49.8) L. lineri Oaxaca El Pacífico Alemania L. smithii Veracruz Guerrero Oaxaca Puerto Escondido Tehuantepec Tapanatepec Chiapas Tonalá Escuintla Guatemala El Salvador (27.3) 5 26 (2.2) 5 6 (5.5) 6 26 (2.5) 6 2 (9.3) 6 27 (2.) 6 22 (8.9) 7 27 (2.3) 7 24 (2.8) 6 24 (8.7) (23.5) (6.7) 5 2 (7.8) 7 8 (7.5) 5 2 (7.4) 5 9 (6.8) 6 22 (7.9) 7 8 (7.4) 6 22 (8.) 5 2 (7.4) 6 8 (7.) (53.6) (52.) (46.5) (5.) (54.4) (5.9) (52.3) (5.8) 46 6 (52.9) 46 5 (48.3) (8.5) (85.6) (9.4) (8.) (87.) 9 23 (2.8) (22.7) (24.) (22.5) (98.3) (23.) Median Prefrontal Width (RMW2). Ratio of the width of the median prefrontal (Fig. : 6 to 2) divided by height of the second postorbital supralabial (Fig. : 28 to 29). Specimens lacking a median prefrontal were scored as. Nasal Length (RNL). Ratio of the length of the midline suture of the nasals (Fig. : to 2) divided by length of the midline suture of the postparietals (Fig. : 5 to 7). Anomalous Postparietal Suture Length (RAPPSL). Ratio of the total length of all anomalous sutures on both postparietals (Fig. : 3 to 4 plus 6 to 8) divided by length of the midline suture of postparietals (Fig. : 5 to 7). Postnasal Height (RPNH). Ratio of the height of the postnasal (Fig. : 22 to 23) divided by height of the anterior loreal (Fig. : 22 to 24). Postocular Length (RPOL). Ratio of the maximum length of the postoculars (Fig. : 26 to 27) divided by length of the orbit (Fig. : 25 to 26). Supralabial Height (RSLH). Ratio of the height of the second postorbital supralabial (Fig. : 28 to 29) divided by height of the first postorbital supralabial (Fig. : 3 to 3). The first postorbital supralabial is usually the sixth and is the first labial scale posterior to the eye that does not enter the orbital ring of scales. Preanal Width (RPAW). Ratio of the maximum width of the lateral posterior preanal (Fig. 3: S) divided by maximum width of the median posterior preanal (Fig. 3: R). AUXILIARY CHARACTERS The following 5 additional characters were used for comparisons among the samples in one of the clusters (group S). Median Prefrontal (M). Presence () or absence () of a median prefrontal. Median Prefrontal Contact (MC). Presence () or absence () of a contact between the median prefrontal and the frontal. Specimens lacking a median prefrontal were scored as. Median Prefrontal Length 2 (RML2). Ratio of the length of the median prefrontal (Fig. : 5 to 8) divided by length of the midline suture of postparietals (Fig. : 5 to 7). The character was scored as absent in specimens lacking a median prefrontal. Prefrontal Length 2 (RPFL2). Ratio of the length along the midline of the lateral prefrontal (Fig. : 4 to 6) divided by length of the midline suture of the nasals (Fig. : to 2). Frontal Length 2 (RFL2). Ratio of the length of the midline suture of frontals (Fig. : 8 to 9) divided by length of the midline suture of postparietals (Fig. : 5 to 7). RESULTS AND DISCUSSION ANALYSIS OF VARIATION AND DELINEATION OF SPECIES Southern Groups An initial discriminant function analysis was performed with the 3 basic meristic and mensural scale characters (excluding the presence/absence of the parietal spot; Table ) for all 29 locality samples from southern México (below 9 N) and Central America. The analysis revealed three nonoverlapping multivariate clusters (groups S, T, and F), that each contain multiple locality samples (Fig. 4). For these three groups, variation in individual characters among the samples is analyzed, the diagnosable units are identified, and species names are assigned. In addition to these three multilocality groups, the initial discriminant analysis identified three clusters (D, L, and R), which each contain a single locality sample. Previous work (Bezy and Camarillo, 992, 997) demonstrated that each of the three samples represents a morphologically discrete species (L. dontomasi Smith, 942; L. lowei Bezy and Camarillo, 997; and L. radula Smith,

11 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma 9 Table 2. Extended. RPAW M MC RML2 RPFL2 RFL2.4.7 (.62) (2.53).6.89 (.77).8.87 (.84).56 (.5).4.2 (.8) (2.2) (.44) (.9).4. (.68) (.64).4. (.65).53.9 (.74).33. (.63).44. (.73) (.62) (.65).47.7 (.57) (.47) (.33) (.2) (.99) (.99) (.8) (.2) (.36) (.37) (.95) (.3).26.8 (.53) (.56) (.5).4.99 (.56).48.8 (.74).3.54 (.44) (3.45). 4.5 (2.45) (3.5). 3.5 (2.33) (2.43) (3.55) (3.7) (3.59) (3.58) (3.6) (.5).23.7 (.79) (.57) (.75) (.86) (.57) (.56) (.58).33.2 (.62) (.46) 942, respectively), and these analyses are not repeated here. However, all species of Lepidophyma recognized as valid are included in the summary and key. GROUP S. The group contains 37 specimens from 33 localities on the Pacific slope from El Salvador to Michoacán (see Specimens Examined for a list of the names of Mexican states that are used in the text without reference to country). Members of the group are characterized by the absence in adults (over 75 mm SVL) of an externally visible parietal spot (PS; Table ). Discriminant function analysis of variation in the 3 characters among the samples of group S revealed that the specimens from Michoacán and Guerrero each constitute a discrete cluster (S and S2, respectively), with a third cluster (S3) composed of multivariately overlapping samples from Oaxaca to El Salvador (Fig. 5). The Michoacán cluster (S) contains 4 specimens from 2 localities. Univariate comparisons among samples in group S indicate that the specimens from Michoacán have discrete (nonoverlapping) differences in number of dorsals and gulars from all oth- Figure 4 Plots of individual specimen scores on the first two canonical variates for 29 locality samples of Lepidophyma from southern México and Central America. Envelopes enclose the specimens for the six nonoverlapping groups present. Figure 5 Plots of individual specimen scores on the first two canonical variates for locality samples of group S of Lepidophyma (from Fig. 4). Three nonoverlapping groups (S, S2, and S3) are present.

12 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma Table 3. Variation in 2 scale characters among samples of group T of Lepidophyma. The observed limits of variation are followed by the mean (in parentheses). N FPT LTR IWD2 PTMP PVTL L. mayae Guatemala (32.3) (.83) 3 4 (3.9) 2 (.9) 2 46 (34.5) L. pajapanense Veracruz (.33) (4.) 3 4 (3.) 6 (7.5) 39 5 (43.5) L. tuxtlae Veracruz Los Tuxtlas Jésus Carranza Oaxaca Vista Hermosa La Gloria Matias Romero (24.5) (23.5) (34.) (36.5) (3.6) (4.2) (5.5) (5.7) (39.2) (39.3) 46 Chiapas Ocote (25.5) 3 33 (3.5) 3 4 (3.8) 4 6 (4.8) 4 46 (43.3) er members of the group (Table 2). The cluster contains the holotype of L. tarascae Bezy, Webb, and Álvarez, 982, and this name is assigned to these specimens. The Guerrero cluster (S2) contains 7 specimens from locality, and these have a higher average number of femoral pores than other members of group S (Table 2). However, the observed limits of variation for femoral pores in the Guerrero sample (S2) slightly overlap those of specimens from Oaxaca and Chiapas (S3), and thus clusters S2 and S3 are judged to be conspecific. Comparisons among the samples of S2 and S3 reveal that 2 localities appear to share a scale feature that is nonoverlapping with other members of Figure 6 Number of longitudinal rows of ventral scales bearing keels plotted against snout vent length for specimens of Lepidophyma lineri (filled circles) and L. smithii (vertical lines). the group. The two individuals from Cafetal El Pacífico (LACM 34844, UCM 48452) and one damaged specimen from the adjacent locality of Cafetal Alemania (AMNH 5977) in the Sierra Madre del Sur of Oaxaca differ in number of lateral tubercle rows from all other specimens of clusters S2 and S3 (Table 2). The 3 specimens include the holotype of L. flavimaculatum lineri Smith, 973, and they are recognized as a distinct species, L. lineri. Smith (973) originally described this taxon as a subspecies of L. flavimaculatum, characterized by the following features (abbreviations in parentheses denote equivalent characters in Table 2): a minute median prefrontal (RML2), subtriangular lateral prefrontals (RPFL2), a large frontal (RFL2), and large lateral preanals (RPAW). The 2 specimens of L. lineri from the type locality appear to have average differences from other samples of clusters S2 and S3 in these 4 characters, but there is extensive overlap in the observed limits of variation (Table 2). As noted by Smith (973) the diagnostic value of the presence of keeled ventrals in this taxon is questionable because the known specimens have a small snout vent length and this feature is negatively correlated with body size (Fig. 6). Although the original diagnostic features used by Smith (973) appear to be overlapping with other samples, the 2 specimens from the type locality (El Pacífico, Oaxaca) have a distinctly higher number of lateral tubercle rows, a feature shared with the specimen from the adjacent locality, Cafetal Alemania (Table 2). Lepidophyma lineri is here recognized as a separate species on the basis of its discrete difference in number of lateral tubercle rows from other members of group S, but this conclusion must be considered tentative because the taxon is represented by only 3 specimens. Although significant differences exist among some of the other samples in S2 and S3, there is

13 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma Table 3. Extended. FTL FTLD DOR DAPVR MC RML RPAW (25.) 9 (4.8) (74.8) 72 9 (79.7) (.8) (2.49) (.5) 26 3 (28.5) 8 (4.2) (65.) (7.) (.32). 2.8 (.73) (.54) (26.) (6.6) (73.5) (8.2) 76 (.89) (.32) (.83) (23.3) (3.8) (59.2) (76.3) 77 (.67) (.34) (.8) (26.5) 4 6 (4.8) (7.) 73 8 (77.8) (.25)..66 (.9) (.63) overlap in the observed limits of variation for all scale characters examined and the remaining samples in these clusters are considered to represent a single species. Cluster S3 includes the holotypes of L. smithii Bocourt, 876; Akleistops guatemalensis Muller, 877; and L. smithii tehuanae Smith, 942. The oldest available name, L. smithii, is assigned to the remaining members of clusters S2 and S3. Smith (942, 973) distinguished L. flavimaculatum tehuanae from L. flavimaculatum smithii by its lack of a median prefrontal. The presence (M) and size (RML2) of the median prefrontal and the presence of a contact between the median prefrontal and the frontal (MC) appear to vary geographically (Table 2). A median prefrontal is present in 33% of the specimens from the vicinity of Tehuantepec, Oaxaca, which encompasses the type locality Figure 7 Plots of individual specimen scores on the first two canonical variates for locality samples of group T of Lepidophyma (from Fig. 4). Three nonoverlapping groups (T, T2, and T3) are present. of L. f. tehuanae. No discrete differences between L. f. tehuanae and L. smithii are apparent in any of the scale characters examined in this study (Table 2) and the two are considered conspecific. The relationship of L. smithii to L. flavimaculatum is discussed under group F, below. On the basis of the above comparisons it is concluded that group S contains 3 morphologically discrete species: L. tarascae (Michoacán), L. lineri (two localities in the Sierra Madre del Sur, Oaxaca), and L. smithii (Pacific slope from Guerrero to El Salvador). GROUP T. The group contains 36 specimens from 3 localities from Veracruz and Oaxaca to Guatemala. Discriminant function analysis of variation in the 3 characters indicates that the samples form three discrete clusters (T, T2, and T3; Fig. 7). Cluster T is composed of 6 specimens from 4 localities in Guatemala. Univariate comparisons identify no discrete differences among the 4 locality samples. Cluster T differs from all specimens of clusters T2 and T3 in pretympanics and from L. smithii, L. lineri, and L. tarascae in lateral tubercle rows (Table ). The cluster contains one holotype, that of L. mayae Bezy, 973, and this name is applied to the species. Cluster T2 contains 22 specimens from 3 regions in Veracruz. Univariate comparisons identify no discrete differences among the samples from these regions. The specimens in cluster T2 differ from all individuals of cluster T3 in femoral pores and fourth toe lamellae and from L. smithii, L. lineri, and L. tarascae in lateral tubercle rows and large paravertebrals (Table ). The cluster contains the holotype of L. pajapanense Werler, 957, and this name is applied to the species. Cluster T3 includes 98 specimens from 6 localities in Veracruz, Oaxaca, and Chiapas. Univariate comparisons among the six samples do not reveal

14 2 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma Figure 8 Number of longitudinal rows of ventral scales bearing keels plotted against snout vent length for specimens of Lepidophyma tuxtlae from the Ocote region of Chiapas (filled circles) and the Tuxtlas region of Veracruz (vertical lines). Figure 9 Plots of individual specimen scores on the first two canonical variates for locality samples of group F of Lepidophyma (from Fig. 4). Three nonoverlapping groups (F, F2, and F3) are present. any discrete differences. The specimens in cluster T3 differ from L. smithii, L. lineri, and L. tarascae in lateral tubercle rows and large paravertebrals. The cluster includes the holotypes of L. tuxtlae Werler and Shannon, 957; L. sawini Smith, 973; and L. alvarezi Smith, 973. The oldest available name, L. tuxtlae, is assigned to the members of T3. Smith (973) described L. alvarezi based on a specimen from the Ocote region of Chiapas. It was distinguished from L. tuxtlae by having keeled ventrals, smaller lateral preanals, and a smaller median prefrontal. The 4 specimens (of group T3) examined from the Ocote region have slight average differences from the sample of L. tuxtlae from the Tuxtlas in relative sizes of the median prefrontal and lateral preanals and in number of femoral pores, lateral tubercle rows, pretympanics, large paravertebrals, and dorsals along the paravertebral row, but there is overlap in the ranges of variation for all scale characters examined (Table 3). The number of longitudinal rows of ventrals bearing keels is negatively correlated with snout vent length and only one of the 4 Ocote specimens is outside of the range of variation found in the L. tuxtlae from the Tuxtla region (Fig. 8). Lepidophyma alvarezi is considered conspecific with L. tuxtlae. A second species, L. sawini, was described by Smith based on a specimen from near Vista Hermosa, Oaxaca. It was distinguished from L. tuxtlae by having relatively large keeled scales bordering the paravertebrals medially. Compared to L. tuxtlae from the Tuxtla region, the 6 specimens now available from the Vista Hermosa region have slightly larger (fewer) scales bordering the paravertebrals medially (DAPVR; Table 3), but there is extensive overlap in the observed limits of variation. The Vista Hermosa specimens have virtually the same numbers of femoral pores as L. tuxtlae from the Tuxtlas and the observed limits of variation do not overlap those of L. pajapanense (Table 3). The Vista Hermosa specimens resemble L. tuxtlae (rather than L. pajapanense) from the Tuxtla region in color pattern (Figs. 23, 3). Although there are slight average differences between the Vista Hermosa sample and L. tuxtlae from the Tuxtla region (Table 3) in lateral tubercle rows, dorsal caudal interwhorls, fourth toe lamellae, divided fourth toe lamellae, dorsals, and dorsals along paravertebral row, there is overlap in all characters examined and L. sawini is considered conspecific with L. tuxtlae. Three species are recognized as valid in group T: L. mayae (Guatemala), L. pajapanense (Veracruz), and L. tuxtlae (Veracruz, Oaxaca, and Chiapas). GROUP F. The group contains 57 specimens from 8 localities in Veracruz and Oaxaca to Panamá. Discriminant function analysis with the 3 characters indicates that the samples form three clusters: F, F2, and F3 (Fig. 9). Cluster F contains 4 specimens from locality in Chiapas. The specimens differ from members of clusters F2 and F3, as well as from all other Lepidophyma, in having a greater number of fourth toe lamellae (Table ). All 4 specimens have 2 longitudinal rows of ventrals at midbody (contra Álvarez and Valentín, 988), a condition that is rare within Lepidophyma, but present in one or more individuals of 6 other species of the genus (Table ). The cluster contains the holotype of L. chicoas-

15 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma 3 Figure Plots of individual specimen scores on the first two canonical variates for locality samples of Lepidophyma sylvaticum and L. flavimaculatum. Of the 2 specimens of L. sylvaticum from Veracruz (VC), the more southern one (a) is placed within L. sylvaticum, whereas the more northern one (b) is placed intermediate between the two species. ense Álvarez and Valentín, 988, and this name is applied to the species. Cluster F2 contains 2 specimens from locality (Río de La Venta) in Chiapas. Although the specimens share some distinctive features, there are notable differences between them in several scale characters (Fig. 9) and the 2 are discussed separately. One of the specimens, the holotype of L. lipetzi Smith and Álvarez del Toro, 977 (UCM 5425), has a higher number of large paravertebral tubercles than do all members of cluster F3. In addition, it differs in having a higher number of dorsal scales and in having a pale color pattern from all F3 specimens from Chiapas, including those from the sympatric (Río de La Venta) and nearby (Selva Ocote) localities. It resembles the 4 specimens of L. chicoasense in having 2 longitudinal rows of ventrals and a pale color pattern, but differs in number of fourth toe lamellae. The second F2 specimen (ENEPI 3792) from Río de La Venta resembles the holotype of L. lipetzi in number of lateral tubercle rows and dorsals and in having a pale color pattern, clearly differing from sympatric (La Venta) and nearby (Ocote) specimens of group F3 in these characters. However, it differs from the holotype in having (vs 2) longitudinal ventral rows, the sixth row from the midline being clearly larger than the adjacent tubercles on only one side of the body. It also differs from the holotype in number of large paravertebral tubercles, resembling F3 specimens in this feature. It is tentatively grouped with the holotype in L. lipetzi. The 2 specimens thus assigned to L. lipetzi differ from L. chicoasense in fourth toe lamellae; from L. smithii, L. lineri, L. tarascae, and L. tuxtlae in femoral pores; and from L. mayae and L. pajapanense in gulars (Table ). They differ in color pattern from all members of group F3 and in number of dorsal scales from sympatric (La Venta) and nearby (Ocote) localities for that group. Variation among Central American locality samples of cluster F3 was examined in detail previously (Bezy, 989) and resulted in the recognition of L. reticulatum Taylor, 955 (Pacific slope of Costa Rica) as distinct from L. flavimaculatum Duméril, 85 (Atlantic slope from México to Panamá); these analyses are not repeated here. The discriminant function analyses indicate that all samples of L. flavimaculatum from southern México (except those from Chiapas) cluster with samples from northern Central America (Guatemala to Honduras). The samples from Chiapas broadly overlap both the southern (Nicaragua to Panamá) and northern Central American clusters. No discrete univariate differences are apparent among the samples of L. flavimaculatum from southern México. Lepidophyma flavimaculatum has discrete differences in scalation (Table ) from L. lineri and L. tarascae (femoral pores), L. chicoasense (fourth toe lamellae), and L. tuxtlae and L. pajapanense (large paravertebral tubercles). It differs from sympatric L. mayae in lateral tubercle rows and pretympanics, and from sympatric (Chiapas) L. lipetzi in number of dorsals. Lepidophyma flavimaculatum and L. smithii comprise multivariately discrete clusters (groups F and S in Fig. 4), including the specimens from where their ranges are in close proximity in Oaxaca. Throughout the extensive ranges of the 2 species they differ in lateral tubercle rows. In addition, L. flavimaculatum has an externally visible pale parietal spot that is absent in adult L. smithii (SVL greater than 75 mm). The 2 also differ in karyotype (Bezy, 972) and are recognized as discrete species. Four species are recognized as valid in group F: L. chicoasense (Chiapas), L. lipetzi (Chiapas), L. reticulatum (Costa Rica), and L. flavimaculatum (Veracruz and Oaxaca to Panamá). Comparisons with Species in Northeastern México Bezy (984) recognized 4 species of Lepidophyma as valid in northeastern México: L. gaigeae Mosauer, 936; L. micropholis Walker, 955; L. occulor Smith, 942; and L. sylvaticum Taylor, 939. Lepidophyma micropholis differs in number of dorsals (Table ) from all southern species delineated above. Lepidophyma gaigeae also differs in dorsals from all southern species except L. radula, from which it differs in femoral pores. Lepidophyma occulor differs in gulars from all southern species except L. smithii, from which it differs in having a parietal spot in adults and a different karyotype. Lepidophyma sylvaticum has discrete differences (Table ) from L. tarascae and L. lineri (femoral pores), L. tuxtlae (dorsals separating paravertebral

16 4 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma tubercles), L. pajapanense (pretympanics), L. chicoasense (fourth toe lamellae), L. smithii (presence of a pale parietal spot in adults), L. lipetzi (height of second postorbital infralabial), and L. mayae (median prefrontal width). Lepidophyma sylvaticum and L. flavimaculatum form discrete multivariate clusters (Fig. ). Where their ranges are in close proximity in Veracruz, the 6 specimens of L. flavimaculatum and the southernmost specimen of L. sylvaticum (Fig., point a; MZFC 722 from Alvarado) are placed well within their respective species clusters. However, the more northern Veracruz specimen of L. sylvaticum (Fig., point b; KU 2699 from Tlapacoyam) is placed somewhat intermediate between the 2 species clusters. Nevertheless, the 2 L. sylvaticum and 6 L. flavimaculatum from Veracruz are nonoverlapping in 5 scale characters (femoral pores, lateral tubercle rows, caudal interwhorls, pretympanics, gulars contacting the first infralabials, and height of the second postorbital supralabial). For the 545 specimens examined from throughout the ranges, the 2 species generally differ in number of pretympanics and in the relative height of the second postorbital supralabial. All but one L. flavimaculatum (UCM from Chiapas) have 4 or more pretympanics and all but 2 L. sylvaticum (AMNH 7273, UMMZ 374, both from Tamaulipas) have 3 or fewer pretympanics. For the relative height of the second postorbital supralabial, all specimens of L. sylvaticum have a RSLH of.84 or greater, and all but 3 specimens of L. flavimaculatum (UMMZ 3777 from Tobasco, MVZ from Guatemala, AMNH 7589 from Panamá) have a RSLH of less than.84. No specimens overlap in both characters. The 2 also have consistent differences in the number and morphology of chromosomes (Bezy, 984) and are considered to represent discrete species. SUMMARY OF THE EXTANT SPECIES OF THE GENUS LEPIDOPHYMA Lepidophyma A. Duméril Lepidophyma A. Duméril, in Duméril and Duméril, 85:37. Type species: Lepidophyma flavimaculatus A. Duméril. Poriodogaster Smith in Gray, 863:54, Plate XXI. Type species: Poriodogaster grayii Smith. Akleistops Muller, 877:39, Tafel I, II. Type species, Akleistops guatemalensis Muller. Gaigeia Smith, 939:24. Type species: Lepidophyma gaigeae Mosauer. DIAGNOSTIC CHARACTERS. The living members of the genus Lepidophyma differ from Cricosaura typica (condition in parentheses) in having frontonasal (2), 2 frontals (), and 2 parietals (none); and from living members of the genus Xantusia (condition in parentheses) in having a round pupil (vertically elliptical); in having the nostrils placed in the suture between the nasal and postnasal (in the suture between the nasal, postnasal, and rostral); and no postmentals (postmentals distinct from infralabials; Savage, 963; R.L. Bezy and J.L.Camarillo, personal observatons). CONTENT. A total of 7 extant species of Lepidophyma are recognized as a result of this study: L. chicoasense, L. dontomasi, L. flavimaculatum, L. gaigeae, L. lineri, L. lipetzi, L. lowei, L. mayae, L. micropholis, L. occulor, L. pajapanense, L. radula, L. reticulatum, L. smithii, L. sylvaticum, L. tarascae, and L. tuxtlae. The validity of L. arizeloglyphum (Langebartel, 953) and other fossil taxa (Estes, 983) not included in this study requires additional investigation focusing on osteology. KARYOTYPES. The diploid chromosome numbers of species in the genus range from 32 to 38; numbers of macrochromosomes, from 6 to 8; and numbers of microchromosomes, from 6 to 2 (Bezy, 972; Bezy and Camarillo, 992, 997). The karyotypes of members of Lepidophyma differ from those of members of Xantusia (Bezy, 972; Bezy and Flores Villela, 999) in having lower diploid numbers (32 38 vs 4) and fewer microchromosomes (6 2 vs 22); and from that of Cricosaura typica (Hass and Hedges, 992) in having higher diploid numbers (32 38 vs 24), and higher numbers of macrochromosomes (6 8 vs 2) and microchromosomes (6 2 vs 2). SIZE. The maximum snout vent length of the 2 species for which or more individuals were examined ranges from 56 mm (L. dontomasi, N 23) to 27 mm (L. flavimaculatum, N 432). DISTRIBUTION, HABITAT, AND NATURAL HISTORY. Extant species of Lepidophyma are found from Nuevo León and Michoacán, México, to Panamá, (Figs., 2), excluding most of the Yucatán peninsula (where the extinct L. arizeloglyphum was found). The greatest species diversity occurs in the region of Oaxaca and Chiapas, México ( species), and in the ranges and valleys associated with the northern Sierra Madre Oriental in Querétaro and San Luis Potosí, México (4 species). Members of the genus occur in a broad range of vegetation types from semiarid thorn scrub to tropical wet forest, cloud forest, and pine and pine oak woodland. The lizards are found under a variety of cover, including rocks, logs, tree stumps, bark, and forest litter; in rock crevices, caves, culverts, and walls; and within human dwellings. Both diurnal and nocturnal activity have been observed, and diet includes fruit and arthropods (Campbell, 998; Mautz and López-Forment, 978). The 5 species for which reproductive information is available are viviparous. REMARKS. Lepidophyma dontomasi, L. gaigeae, L. lowei, and L. radula form a distinct phenetic cluster within the genus characterized by small body size, absence of microtubercles on the scale surfaces (Bezy and Peterson, 988); absence (or small size) of tubercular scales on the body (except L. radula); poorly differentiated caudal whorls and interwhorls; and relatively large (and thus few)

17 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma 5 Figure Geographic distribution of species of Lepidophyma. dorsal, ventral, and gular scales (Bezy and Camarillo, 992, 997). Whether the 4 species comprise a clade (Gaigeia; Smith, 939, 973) or a grade is being addressed by phylogenetic analyses of DNA sequences combined with morphology (R.L. Bezy and collaborators, unpublished). ETYMOLOGY. The name Lepidophyma (Greek: lepidos, scale; phyma, enlarged or swollen object) is neuter (Smith, 973) and refers to the presence of enlarged tubercular scales. COMMON NAMES. Tropical night lizards; lagartijas nocturnas tropicales; escorpiones nocturnos. Lepidophyma chicoasense Álvarez and Valentín Group F; Figure 3 Lepidophyma chicoasensis Álvarez and Valentín, 988:25, Fig.. Holotype: IPN 32; type locality, 6.3 Km N,.6 Km E de Tuxtla Gutiérrez, 6 m., Chiapas, México. DIAGNOSTIC CHARACTERS. The species differs from all other Lepidophyma in having 35 or more fourth toe lamellae. It differs from all except L. smithii, L. flavimaculatum, L. micropholis, and L. sylvaticum in having 53 or more gulars; from L. gaigeae, L. dontomasi, L. lowei, L. radula, L. tarascae, and L. pajapanense in having 77 or more dorsals; and from L. reticulatum, L. micropholis, and L. occulor in having 92 or fewer dorsals (Table ). KARYOTYPE. Unknown. SIZE. The maximum snout vent length for the 4 specimens examined is 78 mm. SEX RATIO. Of the 3 specimens sexed, 2 (67%) are males. DISTRIBUTION, HABITAT, AND NATURAL HISTORY. The species is known only from the type locality located in the Cañón Sumidero (Fig. ). REMARKS. Lepidophyma chicoasense resembles L. lipetzi in color pattern and in many features of scalation, including the presence of 2 longitudinal rows of ventrals at midbody (contra Álvarez and Valentín, 988). However, it has a higher number of fourth toe lamellae than all other members of the genus, including L. lipetzi, L. flavimaculatum, and L. tuxtlae, which occur in the region. ETYMOLOGY. The name chicoasense is an adjective and refers to La Presa de Chicoasén, Chiapas, located in the region of the type locality. The original name is emended in this paper to modify the neuter generic name. COMMON NAMES. Sumidero Tropical Night Lizard; Lagartija Nocturna del Sumidero.

18 6 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma Figure 2 Geographic distribution of seven species of Lepidophyma. Lepidophyma dontomasi (Smith) Group D; Figure 4 Gaigeia dontomasi Smith, 942:374. Holotype: USNM 473; type locality, Lachiguirí, Oaxaca, at 7, feet. Lepidophyma dontomasi: Savage, 963:35. DIAGNOSTIC CHARACTERS. The species differs from all other Lepidophyma except L. lowei and L. radula in having enlarged caudal whorls that are separated dorsally by 2 rows of interwhorls, only one of which is complete ventrally; from all except L. gaigeae, L. radula, and L. tarascae in having 49 or fewer dorsals; from L. gaigeae in having 2 or fewer femoral pores; from L. tarascae in having 35 or fewer gulars (Table ); and from L. radula in lacking vertical rows of distinctly enlarged tubercles on the side of the body. KARYOTYPE. Lepidophyma dontomasi has a diploid chromosome number of 32 with 6 macrochromosomes and 6 microchromosomes (Bezy and Camarillo, 992). The macrochromosomes are very similar to those of L. smithii. The species has the lowest number of total chromosomes and of microchromosomes in the genus. SIZE. The maximum snout vent length for the 29 specimens examined is 56 mm, smaller than any other member of the genus except L. radula (known only from the holotype). SEX RATIO. Of the 22 specimens sexed, 9 (4%) are males. DISTRIBUTION, HABITAT, AND NATURAL HISTORY. The species is known only from the type locality on Cerro Lachiguirí, Oaxaca (Fig. ), where it has been found under logs in pine forest (Smith, 942) and in rock crevices and under rocks and dead Agave and Nolina in oak savanna (Bezy and Camarillo, 992). REMARKS. Additional field work is needed to determine if the species occurs elsewhere in the Sierra de los Mijes. ETYMOLOGY. The name dontomasi is a noun in the genitive singular case and honors Thomas (Don Tomas) Baillie MacDougall ( ), a naturalist who collected plants and animals (including the holotype of L. dontomasi) from remote areas in México, particularly the Tehuantepec region (Smith, 974). COMMON NAMES. MacDougall s Tropical Night Lizard; Lagartija Nocturna de MacDougall. Lepidophyma flavimaculatum complex Group F3; Figure 5 Lepidophyma flavimaculatus A. Duméril in Duméril and Duméril, 85:38. Holotype: MNHN 782; type locality, Province du Petén (America centrale) (restricted to Río de la Pasión, Guatemala, by Smith and Taylor, 95).

19 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma 7 Figure 3 The holotype of Lepidophyma chicoasense (IPN 32; snout vent length, 78 mm). Figure 4 Living individual of Lepidophyma dontomasi (ENEPI 36; snout vent length, 5 mm).

20 8 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma Figure 5 Living individual of Lepidophyma flavimaculatum (ENEPI 3793; snout vent length, 7 mm). Lepidophyma flavimaculatum: A.H.A. Duméril, 852, Plate 7. Poriodogaster grayii Smith in Gray, 863:54, Plate XXI. Holotype: BMNH ; type locality, unknown (originally conjectured to be Lower California and subsequently thought by Peters, 874, to be Panamá). Lepidophyma flavomaculatum obscurum Barbour, 924:. Holotype: MCZ 7747; type locality, Río Chilibrillo, Panamá. Lepidophyma flavomaculatum flavomaculatum: Smith, 942:379. Lepidophyma anomalum Taylor, 955:554, Fig. 4. Holotype: KU 3427; type locality, Los Diamantes, Guápiles, Limón Province, Costa Rica. Lepidophyma ophiophthalmum Taylor, 955:558, Fig. 5. Holotype: KU 3625; type locality, 5 km. NNE Tilarán, Guanacaste, Costa Rica. DIAGNOSTIC CHARACTERS. The species differs from L. dontomasi, L. radula, L. tarascae, L. lineri, and L. occulor in having 25 or more femoral pores; from L. dontomasi, L. gaigeae, L. radula, and L. tarascae, in having 7 or more dorsals; from L. dontomasi, L. lowei, and L. radula in having 4 or more gulars; from L. pajapanense and L. tuxtlae in having 27 or fewer large paravertebral tubercles; from L. smithii in having 23 or more lateral tubercle rows; from L. micropholis in having 225 or fewer dorsals; from L. chicoasense in having 3 or fewer fourth toe lamellae; from L. lipetzi in having 8 or more dorsals (99.3% of all L. flavimaculatum, % of Chiapas specimens) and a darker color pattern (Fig. 5); from L. reticulatum in lacking a boldly reticulated color pattern on the gular surface; from L. mayae in having 33 or fewer lateral tubercle rows (99.6% of all L. flavimaculatum, % of Guatemala specimens) and 3 or more pretympanics (99.8% of all L. flavimaculatum, % of Guatemala specimens); from L. sylvaticum in having 4 or more pretympanics (99.3% of all L. flavimaculatum) and a lower second postorbital supralabial (.83 or less; 99.3% of all L. flavimaculatum: Table ); and from L. reticulatum in lacking a bold reticulations on the gular surface. KARYOTYPE. Lepidophyma flavimaculatum has a diploid chromosome number of 38 with 8 macrochromosomes and 2 microchromosomes (Bezy, 972). One individual from an all-female population appeared to be a diploid triploid mosaic. The karyotype is identical to that of L. pajapanense, L. tuxtlae, and L. reticulatum. It differs from that of L. sylvaticum in having 8 rather than 6 microchromosomes. SIZE. The maximum snout vent length for the 432 specimens examined is 27 mm.

21 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma 9 SEX RATIO. Of the 38 specimens sexed, 6 (9%) are males. No males were present in samples from Panamá and most populations in Costa Rica. DISTRIBUTION, HABITAT, AND NATURAL HISTORY. The species is found primarily on the Atlantic versant from Veracruz and Oaxaca to Panamá, but absent from the outer Yucatán Península (Fig. 2). It occurs in sympatry with L. pajapanense in Veracruz; with L. tuxtlae in Veracruz, Oaxaca, and Chiapas; with L. lipetzi in Chiapas; with L. mayae in Guatemala; and with L. reticulatum in Costa Rica. The species occurs at low and moderate elevations in tropical moist forest, tropical dry forest, and subtropical wet forest, where it is found in rotting logs and tree stumps; in forest litter; under bark; under rocks; and in rock crevices, caves, and ruins (Álvarez del Toro, 982; Burt, 935; Campbell, 998; Duellman, 963; Henderson and Hoevers, 975; Lee, 996; Meyer and Wilson, 973; Scott, 976; Slevin, 942; Stuart, 948; Telford and Campbell, 97; R.L. Bezy and colleagues, personal observations). It is reported to consume termites, ants, crickets, spiders, scorpions, centipedes, and millipedes, and to be preyed upon by the colubrid snakes Dryadophis melanolomus Cope, 868, and Drymarchon corais Boie, 827 (Campbell, 998; Lieberman, 986). In Chiapas, five to eight young are born in June and July (Álvarez del Toro, 982), and in Panamá, a female gave birth in June to five neonates, all females (Telford and Campbell, 97). REMARKS. Lepidophyma flavimaculatum is unusually variable in scalation because of geographical variation among its widely scattered localities. It is most similar to L. mayae, L. lipetzi, and L. sylvaticum. In areas where it occurs with L. mayae and L. lipetzi, individual scale characters segregate all specimens, indicating reproductive isolation. Univariate scale characters separate 99% of the specimens of L. flavimaculatum from L. sylvaticum, and the 2 species differ in karyotype for all localities examined (L. flavimaculatum from Chiapas to Panamá and L. sylvaticum from Nuevo León to Hidalgo). The term complex is used for the aggregation of Lepidophyma flavimaculatum populations to denote that it may contain more than species. The populations in Panamá and most of Costa Rica lack males but do not appear to differ significantly in allozymes, karyotype, scalation, or color pattern from the populations to the northwest that contain males (Bezy, 972, 989; Bezy and Sites, 987). ETYMOLOGY. The name flavimaculatum (Latin: flavus, yellow; maculatum, spotted) is an adjective and refers to the dorsal color pattern. The original name was justifiably emended by A.H.A. Duméril (852) to modify the neuter generic name. COMMON NAMES. Yellow-spotted Tropical Night Lizard; Lepidofima; Escorpión Nocturno Puntos Amarillos; Reina de Culebra; Perrozompopo Atlántico. Lepidophyma gaigeae Mosauer Group G; Figure 6 Lepidophyma gaigeae Mosauer, 936:3, Plate II. Holotype: MCZ 4245 (Mosauer Mexican Collection 97); type locality, Durango, State of Hidalgo, México. Gaigeia gaigeae: Smith, 939:24. DIAGNOSTIC CHARACTERS. The species differs from all other Lepidophyma except L. tarascae in having enlarged caudal whorls that are separated dorsally by 2 rows of interwhorls both of which are complete ventrally; from all except L. dontomasi, L. radula, L. tarascae, L. tuxtlae, and L. sylvaticum in having 5 or fewer dorsals; from L. dontomasi, L. radula, and L. tarascae in having 28 or more femoral pores; from L. tuxtlae in lacking paravertebral rows composed of a continuous string of subequal tubercles; and from L. sylvaticum in lacking distinctly enlarged tubercles on the side of the body (Table ). KARYOTYPE. Lepidophyma gaigeae has a diploid chromosome number of 38 with 8 macrochromosomes and 2 microchromosomes (Bezy, 972). The karyotype is similar to that of L. flavimaculatum, L. tuxtlae, and L. pajapanense, but differs in centromere position for 2 pairs of chromosomes. SIZE. The maximum snout vent length for the 77 specimens examined is 66 mm. SEX RATIO. Of the 47 specimens sexed, 55 (37%) are males. DISTRIBUTION. The species occurs in Querétaro and Hidalgo (Fig. ), where it is found primarily in limestone crevices in pine oak woodland along the Sierra Madre Oriental (Dixon et al., 972; Taylor, 939; R.L. Bezy and colleagues, personal observations). It has also been found on an adobe wall in thorn scrub in the Jalpan Valley, Querétaro (Dixon et al., 972), and in a cave in xerophytic scrub near the Río Tula, Hidalgo (González, 995). It occurs in sympatry with L. sylvaticum and L. occulor in Querétaro. A captive female gave birth to one young in March (Dixon et al., 972). The species appears to be diurnal based on metabolic data (Mautz, 979) and it is more resistant to water loss than is L. smithii (Mautz, 982). REMARKS. The populations found in the Jalpan Valley differ slightly in scalation (Bezy and Camarillo, 992). ETYMOLOGY. The name gaigeae is a noun in the genitive singular case and honors Helen Thompson Gaige (89 976), former curator at the University of Michigan Museum of Zoology. COMMON NAMES. Gaige s Tropical Night Lizard; Lagartija Nocturna de Gaige. Lepidophyma lineri Smith Group S; Figure 7 Lepidophyma flavimaculatum lineri Smith, 973: 8, Plates 2. Holotype: UCM 48452; type lo-

22 2 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma Figure 6 Living individual of Lepidophyma gaigeae (ENEPI 488; snout vent length, 63 mm). cality, Cafetal Pacífico, 2. mi. N and.4 mi. E Candelaria Loxicha, Oaxaca, on the trail to Cafetal Hedweges (about 7 km. N Pochutla). DIAGNOSTIC CHARACTERS. The species differs from all other Lepidophyma except L. smithii and L. tarascae in lacking a pale parietal spot; from L. smithii in having 23 or more lateral tubercle rows; from L. tarascae in having 79 or more dorsals and 5 or more gulars; from L. occulor in having 53 or fewer gulars and 84 or fewer dorsals; from all except L. dontomasi, L. smithii, L. tarascae, and L. occulor in having 7 or fewer total femoral pores; and from all except L. radula, L. smithii, L. tarascae, L. flavimaculatum, L. occulor, and L. sylvaticum in having 24 or fewer lateral tubercle rows (Table ). KARYOTYPE. Unknown. SIZE. The maximum snout vent length for the two undamaged specimens examined is 37 mm. SEX RATIO. Both specimens sexed are males. DISTRIBUTION, HABITAT, AND NATURAL HISTORY. The species is known only from the type locality and nearby Cafetal Alemania in the Sierra de Miahuatlán of Oaxaca (Fig. 2). REMARKS. Lepidophyma lineri is most similar to L. smithii. Additional material is needed to clarify its variation, distribution, and geographic relationship with L. smithii, which occurs ca. 8 km to the west along road to Puerto Escondido, Oaxaca. ETYMOLOGY. The name lineri is a noun in the genitive singular case and honors Ernest A. Liner, one of the herpetologists who collected the holotype. COMMON NAMES. Liner s Tropical Night Lizard; Lagartija Nocturna de Liner. Lepidophyma lipetzi Smith and Álvarez del Toro Group F2; Figure 8 Lepidophyma lipetzi Smith and Álvarez del Toro, 977:37, Figs Holotype: UCM 5425; type locality, Lago de Mal Paso, headwaters of Río de La Venta, 3 km N Cintalapa (straight line), Chiapas, México. DIAGNOSTIC CHARACTERS. The species differs from all other Lepidophyma except L. smithii, L. lineri, L. tuxtlae, L. chicoasense, L. flavimaculatum, L. reticulatum, L. micropholis, L. occulor, and L. sylvaticum in having 46 or more gulars; from L. micropholis and L. occulor in having 52 or fewer gulars; from all except L. mayae, L. pajapanense, L. chicoasense, L. flavimaculatum, L. reticulatum, L. micropholis, and L. sylvaticum in having 35 or more femoral pores; from all except L. gaigeae, L. smithii, L. pajapanense, L. tuxtlae, L. chicoasense, L. flavimaculatum, and L. reticulatum in having 6 or more pretympanics; from L. chicoasense in having 28 or fewer fourth toe la-

23 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma 2 Figure 7 The holotype of Lepidophyma lineri (UCM 48452; snout vent length, 36 mm). mellae; from L. reticulatum in having 35 or more femoral pores, 4 or fewer divided fourth toe lamellae, and 79 or fewer dorsals; and from L. flavimaculatum in having 79 or fewer dorsals (99.3% of all L. flavimaculatum, % of Chiapas specimens; Table ) and a pale dorsal color pattern (Fig. 8). KARYOTYPE. Unknown. SIZE. The maximum snout vent length for the two specimens examined is 55 mm. SEX RATIO. Unknown DISTRIBUTION, HABITAT, AND NATURAL HISTORY. The species is known only from the type locality (Fig. ), where it occurs with L. flavimaculatum. The 2 known specimens were collected on the walls of caves (Smith and Álvarez del Toro, 977). REMARKS. The distributional relationship in Chiapas of L. lipetzi, found on the Río de La Venta, and L. chicoasense, known only from the Cañón Sumidero of the Río Grijalva, may never be resolved because the vast Lago Netzahualcóyotl now lies at the confluence of these rivers. ETYMOLOGY. The name lipetzi is a noun in the genitive singular case and honors Milton L. Lipetz, a University of Colorado faculty member admired by Hobart Smith. COMMON NAMES. Lipetz s Tropical Night Lizard; Lagartija Nocturna del Ocote. Lepidophyma lowei Bezy and Camarillo Group L; Figure 9 Lepidophyma lowei Bezy and Camarillo, 997:, Figs. 3. Holotype: IBH 75; type locality, 4. km (by rd) SE San Bartolomé Zoogocho, Municipio Zoogocho, former Distrito Villa Alta, Oaxaca, México (7 4 N, 96 5 W; ca. 22 m elev.). DIAGNOSTIC CHARACTERS. The species differs from all other Lepidophyma except L. dontomasi and L. radula in having enlarged caudal whorls that are separated dorsally by 2 rows of interwhorls, only one of which is complete ventrally; from all except L. dontomasi, L. gaigeae, L. radula, L. tuxtlae, and L. pajapanense in having 37 or fewer gulars; from L. dontomasi, L. gaigeae, and L. radula in having 58 or more dorsals; and from L. tuxtlae, L. pajapanense, L. mayae, L. chicoasense, L. lipetzi, L. flavimaculatum, L. reticulatum, and L. micropholis in having 7 or fewer divided fourth toe lamellae (Table ). KARYOTYPE. Lepidophyma lowei has a diploid chromosome number of 36 with 6 macrochromosomes and 2 microchromosomes (Bezy and Camarillo, 997). The karyotype appears virtually identical to that of L. smithii. SIZE. The maximum snout vent length for the 6 specimens examined is 6 mm. SEX RATIO. Of the specimens sexed, 6 (6%) are males. DISTRIBUTION, HABITAT, AND NATURAL HISTORY. The species is known only from the vicinity of the type locality (Fig. ), where it is found in rock crevices in pine oak woodland. A female gave birth to four offspring in April (Camarillo, 999). REMARKS. Additional work is needed to establish the geographic distribution of the species. ETYMOLOGY. The name lowei is a noun in the genitive singular case and honors Charles H. Lowe, a herpetologist at the University of Arizona. COMMON NAMES. Lowe s Tropical Night Lizard; Lagartija Nocturna de Lowe. Lepidophyma mayae Bezy Group T; Figure 2 Lepidophyma mayae Bezy, 973:, Figs. 3. Holotype: KU 59554; type locality, near Chinajá, elev. 4 m., Depto. Alta Verapaz, Guatemala.

24 22 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma Figure 8 Living individual of Lepidophyma lipetzi (ENEPI 3792; snout vent length, 53 mm). Figure 9 Living individual of Lepidophyma lowei (IBH 75; snout vent length, 53 mm).

25 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma 23 Figure 2 Living individual of Lepidophyma mayae (UTA-R 3822; snout vent length, 78 mm). DIAGNOSTIC CHARACTERS. The species differs from L. radula, L. tarascae, L. lineri, L. pajapanense, L. tuxtlae, L. chicoasense, L. lipetzi, and L. reticulatum in having 2 or fewer pretympanics; from L. dontomasi, L. gaigeae, L. radula, and L. tarascae in having 62 or more dorsals; from L. reticulatum, L. micropholis, and L. occulor in having 88 or fewer dorsals; from L. lowei, L. radula, L. smithii, L. tarascae, L. lineri, L. chicoasense, and L. occulor in having 33 or more lateral tubercle rows; from L. sylvaticum in having a wider median prefrontal (RMW2 greater than.6); and from L. flavimaculatum in having 33 or more lateral tubercle rows and two or fewer pretympanics (99.6% of all L. flavimaculatum, % of Guatemalan specimens; Table ). KARYOTYPE. Unknown. SIZE. The maximum snout vent length for the 6 specimens examined is 75 mm. SEX RATIO. Of the specimens sexed, 4 (4%) are males. DISTRIBUTION, HABITAT, AND NATURAL HISTORY. The specimens examined are from 3 areas in Guatemala: the southern El Petén near the base of the Sierra de Chinajá (border of Deptos. El Petén and Alta Verapaz), the Sierra del los Cuchumatanes (Depto. Huehuetenango), and the valley of the Río Cahabón near the lower slopes of the Sierra Xucaneb (Finca Volcán, Depto. Alta Verapaz; Fig. ). They were found in lowland rain forest (Chinajá, 4 m; Duellman, 963) and in an area supporting cloud forest (Finca Chiblac, ca. 3 m.; Wake and Elias, 983). At Chinajá the specimens were taken within the rain forest under rocks and logs or active on the forest floor during the day (W.E. Duellman, 96 field notes). The species is sympatric with L. flavimaculatum at Chinajá and Finca Volcán. Recently, Campbell (998) recorded the species from additional localities in the Montañas del Mico, Sierra de Santa Cruz, and Maya Mountains north into Belize. The specimens were found in tropical and subtropical wet forest in limestone crevices and under rocks and logs. REMARKS. Although Lepidophyma mayae can be distinguished from all L. flavimaculatum only by a combination of 2 characters (lateral tubercle rows and pretympanics), where the two species are found in sympatry each of the two characters is nonoverlapping, indicating reproductive isolation. ETYMOLOGY. The name mayae is a noun in the genitive singular case and refers to the Maya, the indigenous people of the region of the type locality. COMMON NAMES. Mayan Tropical Night Lizard; Escorpión Nocturno Maya.

26 24 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma Figure 2 Living individual of Lepidophyma micropholis (LACM 6767; snout vent length, 87 mm). Lepidophyma micropholis Walker Figure 2 Lepidophyma micropholis Walker, 955:6. Holotype: UMMZ 298; type locality, cave at El Pachón, about 5 miles NNE of Antigua Morelos, Tamaulipas, estimated elevation 6 7 feet. DIAGNOSTIC CHARACTERS. The species differs from all other Lepidophyma except L. occulor in having 23 or more dorsal scales. It differs from L. occulor, L. dontomasi, L. radula, L. tarascae, and L. lineri in having 28 or more femoral pores; and from all except L. smithii, L. flavimaculatum, L. chicoasense, L. occulor, and L. sylvaticum in having 55 or more gulars (Table ). KARYOTYPE. Lepidophyma micropholis has a diploid chromosome number of 36 with 6 macrochromosomes and 2 microchromosomes. The karyotype appears identical to that of L. sylvaticum. SIZE. The maximum snout vent length for the 3 specimens examined is mm. SEX RATIO. Of the 23 specimens sexed, 2 (52%) are males. DISTRIBUTION, HABITAT, AND NATURAL HISTORY. The species is known from four localities in Sierra del Abra of Tamaulipas and San Luis Potosí (Fig. ) where it is found in limestone caves and crevices. The vegetation near the type locality consists of tropical deciduous forest. Within the caves, the lizards have been observed active during the day near the mouth, where they are found on ledges and in crevices in the walls and mud floor (Walker, 955; R.L. Bezy and colleagues, personal observations). REMARKS. Lepidophyma micropholis appears to be a facultative troglodyte, because a number of individuals have been observed to be active outside caves (J.R. Dixon, personal communication). ETYMOLOGY. The name micropholis (Greek: mikros, small; pholis, scale) is a noun in apposition and presumably refers to the small size of the dorsal scales. COMMON NAMES Cave Tropical Night Lizard; Lagartija Nocturna de Cueva. Lepidophyma occulor Smith Figure 22 Lepidophyma smithii occulor Smith, 942:378. Holotype: USNM 4733; type locality, Jalpan, Querétaro. Lepidophyma flavimaculatum occulor: Walker, 955:5. Lepidophyma occulor: Bezy, 972:5. DIAGNOSTIC CHARACTERS. The species dif-

27 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma 25 Figure 22 Living individual of Lepidophyma occulor (TCWC 3565; snout vent length, 9 mm). fers from all other Lepidophyma except L. smithii and L. micropholis in having 58 or more gulars; from all except L. radula, L. smithii, L. lineri, L. tarascae, L. flavimaculatum, and L. sylvaticum in having 25 or fewer lateral tubercle rows; and from L. smithii in having 22 or more lateral tubercle rows (99%; Table ) and in having a pale parietal spot throughout life. KARYOTYPE. Lepidophyma occulor has a diploid chromosome number of 36 with 8 macrochromosomes and 8 microchromosomes (Bezy, 972). The karyotype is unique within the Xantusiidae. SIZE. The maximum snout vent length for the specimens examined is 5 mm. SEX RATIO. Of the specimens sexed, 7 (64%) are males. DISTRIBUTION, HABITAT, AND NATURAL HISTORY. The species occurs in thorn scrub in the Jalpan valley of Querétaro and San Luís Potosí (Fig. ), where the lizards have been found under rocks or active at night near limestone outcrops and in rock walls and ruins (Dixon et al., 972; R.L. Bezy and J.L. Camarillo, personal observations). East of the Jalpan region, it has been taken in a cave in tropical forest along the Río Moctezuma near Tamazunchale, San Luis Potosí (Camarillo, 993). It occurs in sympatry with L. gaigeae. REMARKS. Lepidophyma occulor is most similar to L. smithii, from which it differs in karyotype and in the presence of a parietal spot. ETYMOLOGY. The name occulor (Latin: one who hides) is a noun in apposition and refers to the secretive or reclusive behavior of the members of the genus (H.M. Smith, personal communication). COMMON NAMES. Jalpan Tropical Night Lizard; Lagartija Nocturna de Jalpan. Lepidophyma pajapanense Werler Group T2; Figure 23 Lepidophyma pajapanensis Werler, 957:223, Plate, Fig.. Holotype: FMNH 78382; type locality, southeast slopes of Volcán San Martín Pajapán, Veracruz, México; elevation 3,5 feet. Lepidophyma pajapanense: Pérez-Higareda, 978: 69. DIAGNOSTIC CHARACTERS. The species differs from all other Lepidophyma except L. tuxtlae in having the paravertebral rows composed of a nearly uninterrupted string of subequal tubercles; from L. tuxtlae, L. dontomasi, L. gaigeae, L. radula, L. tarascae, L. lineri, and L. occulor in having 3 or more femoral pores; from all except L. mayae, L. tuxtlae, and L. sylvaticum in having 39 or more large paravertebral tubercles; and from L. dontomasi, L. lowei, L. radula, L. tarascae, L. lineri, L. mayae, L. micropholis, L. occulor, and L. sylvaticum in having 6 or more pretympanics (Table ) KARYOTYPE. Lepidophyma pajapanense has a diploid chromosome number of 38 with 8 macrochromosomes and 2 microchromosomes (Bezy, 972). The karyotype appears to be identical to that of L. tuxtlae, L. flavimaculatum, and L. reticulatum.

28 26 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma Figure 23 Living individual of Lepidophyma pajapanense (LACM 355; snout vent length, 57 mm). SIZE. The maximum snout vent length for the 22 specimens examined is 82 mm. SEX RATIO. Of the 6 specimens sexed, 5 (3%) are males. DISTRIBUTION, HABITAT, AND NATURAL HISTORY. The species is known only from Veracruz and occurs in the Sierra de Los Tuxtlas (Volcán San Martín Tuxtla to Volcán San Martín Pajapán) and from the vicinity of the Río Coatzacoalcos near Jésus Carranza on the Isthmus of Tehuantepec. It is sympatric with L. tuxtlae in the Tuxtlas region, and with L. tuxtlae and L. flavimaculatum near the Río Coazacoalcos (Fig. ). The locality data for the specimen reported from Huayacacotla, Veracruz (Aguilar Cortés et al., 99), is now considered questionable because extensive collecting in the area has yielded numerous L. sylvaticum, but no additional specimens of L. pajapanense. In the Tuxtlas the lizards are nocturnal and found beneath the bark and between the buttresses of trees within the rain forest from sea level to 5 m (Vogt et al., 997). Both L. pajapanense and L. tuxtlae have been taken from the same tree stump in recent clearings near the forest edge in the Tuxtlas (R.L. Bezy and C.J. Cole, personal observations). The male and female reproductive cycles are asynchronous and four to eight young are born in late May or early June (Méndez-de la Cruz et al., 999). REMARKS. This morphologically distinctive species has a rather restricted distribution, which fortunately includes the preserve at Estación de Biología Tropical Los Tuxtlas. ETYMOLOGY. The name pajapanense is an adjective and refers to the type locality, Volcán San Martín Pajapán. The original name was justifiably emended by Pérez-Higareda (978) to modify the neuter generic name. COMMON NAMES. Pajapan Tropical Night Lizard; Lagartija Nocturna de Pajapán. Lepidophyma radula (Smith) Group R; Figure 24 Gaigeia radula Smith, 942:374. Holotype: USNM 472; type locality, San José Manteca, 5 kilometers from San Carlos Yautepec, Oaxaca. Lepidophyma radula: Savage, 963:35. DIAGNOSTIC CHARACTERS. The species differs from all other Lepidophyma except L. dontomasi and L. lowei in having enlarged caudal whorls that are separated dorsally by 2 rows of interwhorls, only one of which is complete ventrally; from all except L. gaigeae and L. dontomasi in having less than 4 dorsals; from L. dontomasi in

29 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma 27 Figure 24 The holotype of Lepidophyma radula (USNM 472; snout vent length, 4 mm). having distinctly enlarged tubercles on the side of the body; and from L. lowei and L. gaigeae in having fewer than 22 femoral pores (Table ). KARYOTYPE. Unknown. SIZE. The snout vent length of the only known specimen is 4 mm. SEX RATIO. The holotype is a female. DISTRIBUTION, HABITAT, AND NATURAL HISTORY. The species is known only from the type locality (Fig. ), where the vegetation consists of tropical deciduous forest. REMARKS. Lepidophyma radula remains known only from the holotype. Considerable field work conducted by the authors in the region of the type locality has failed to produce additional specimens. Lepidophyma smithii occurs nearby (Portillo Nejapa, Oaxaca) and differs extensively in scalation. Lepidophyma dontomasi, L. lowei, and L. radula are restricted to eastern Oaxaca and are unique in the genus in having most caudal whorls separated by only 2 interwhorls, one of which is complete ventrally. ETYMOLOGY. The name radula (Latin: scraper) is a noun in apposition and refers to the body surface, which is rough, or filelike because of the presence of large tubercular scales (H.M. Smith, personal communication). COMMON NAMES. Yautepec Tropical Night Lizard; Lagartija Nocturna de Yautepec. Lepidophyma reticulatum Taylor Group F3; Figures 25, 26 Lepidophyma reticulatum Taylor, 955:55, Fig. 4. Holotype: KU 36245; type locality, Agua Buena, Puntarenas Province, Costa Rica. Lepidophyma flavimaculatum reticulatum: Wermuth, 965:96. DIAGNOSTIC CHARACTERS. The species differs from all other Lepidophyma except L. smithii, L. flavimaculatum, L. lipetzi, L. micropholis, L. occulor, and L. sylvaticum in having 93 or more dorsals; from L. micropholis in having 229 or fewer dorsals; from L. dontomasi, L. radula, L. tarascae, L. lineri, and L. occulor in having 27 or more femoral pores; from L. lipetzi in having 33 or fewer femoral pores; from L. smithii, L. tarascae, L. lineri, and L. occulor in having 28 or more lateral tubercle rows; from L. sylvaticum in having a lower second postorbital supralabial (RSLH.7 or less; Table ); and from L. flavimaculatum in having a boldly reticulated color pattern on the gular surface (Fig. 26). KARYOTYPE. Lepidophyma reticulatum has a diploid chromosome number of 38 with 8 macrochromosomes and 2 microchromosomes (Bezy, 972, as L. flavimaculatum). The karyotype appears to be virtually identical to that of L. pajapanense, L. tuxtlae, and L. flavimaculatum. SIZE. The maximum snout vent length for the 65 specimens examined is 3 mm. SEX RATIO. No males are present among 29 specimens sexed. DISTRIBUTION, HABITAT, AND NATURAL HISTORY. Lepidophyma reticulatum ranges on the Pacific slope of Costa Rica from the area of Tilarán (Guanacaste Province) to the Península de Osa and San Vito de Java (Puntarenas Province; Fig. 2). In the Tilarán area it occurs in sympatry with a population of L. flavimaculatum that includes males. The species has been collected within and under decaying logs and beneath rocks (R.L. Bezy and K. Bolles, personal observations). REMARKS. This unisexual species may range into southwestern Panamá. ETYMOLOGY. The name reticulatum (Latin: netlike or reticulated) is an adjective and refers to the gular color pattern. COMMON NAMES. Costa Rican Tropical Night Lizard; Lagartija Nocturna de Costa Rica. Lepidophyma smithii Bocourt Groups S2, S3; Figure 27 Lepidophyma smithii Bocourt, 876:42. Syntypes: MNHN 76.95, 4323, A4323, B4323, 4968, A4968; type locality, Tehuantepec and Guatemala occidental (restricted to Guatemala by Smith, 942, and to Mazatenango by Smith and Taylor, 95). Akleistops guatemalensis Muller, 877:39, Tafel I,

30 28 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma Figure 25 Living individual of Lepidophyma reticulatum (LACM; snout vent length, 9 mm). II. Syntypes: NMBA , 3752, 839 4; type locality, Mazatenango (costa grande von Guatemala). Lepidophyma smithii tehuanae Smith, 942:377. Holotype: USNM 488; type locality, Cerro Figure 26 Gular pattern of a living individual of Lepidophyma reticulatum (LACM; snout vent length, 9 mm). Arenal, 3 kilometers west of Tehuantepec, Oaxaca. Lepidophyma smithii smithii: Smith, 942:38. Lepidophyma flavimaculatum smithii: Walker, 955:5. DIAGNOSTIC CHARACTERS. The species differs from all other Lepidophyma except L. tarascae and L. lineri in lacking a pale spot on the parietal scale in adults (snout vent length greater than 75 mm); from all except L. tarascae, L. occulor, and L. sylvaticum in having 22 or fewer lateral tubercle rows; and from L. tarascae in having 62 or more dorsals and 44 or more gulars (Table ). KARYOTYPE. Lepidophyma smithii has a diploid chromosome number of 36 with 6 macrochromosomes and 2 microchromosomes (Bezy, 972). The macrochromosomes are similar to those of L. dontomasi (Bezy and Camarillo, 992), but the 2 species differ in number of microchromosomes (2 vs 6). SIZE. The maximum snout vent length for the 29 specimens examined is 2 mm. SEX RATIO. Of the 234 specimens sexed, 9 (39%) are males. DISTRIBUTION, HABITAT, AND NATURAL HISTORY. The species ranges along the Pacific versant from Guerrero to El Salvador (Fig. 2). One specimen (TCWC 2676) represents a strange geo-

31 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma 29 Figure 27 Living individual of Lepidophyma smithii (LACM 355; snout vent length, 92 mm). graphical anomaly. The specimen bears a field tag with the locality 2 mi NW Sontecomapan, Veracruz. The species is otherwise unrecorded from the Atlantic versant and no other specimens are present among the many individuals of L. tuxtlae and L. pajapanense from the Los Tuxtlas region. The specimen is clearly assignable to L. smithii (based on femoral pores, lateral tubercle rows, and large paravertebrals; Table 2). It has a parietal spot, a feature that is absent in adult L. smithii (Table 2), but the snout vent length (4 mm) of the specimen is within the range of the few juvenile L. smithii in which the spot is visible (28 75 mm). The locality is here considered questionable for L. smithii, but the possibility that the species occurs in the Tuxtlas cannot be completely ruled out in view of the many biogeographic enigmas presented by xantusiids. The species occurs in a variety of habitats. At the northernmost known locality, near Puerto Marqués, Guerrero, it occurs in a system of caves formed by piles of large granitic boulders in tropical semideciduous forest. Within the caves, the lizards have been found to be active day and night in both the dry and wet seasons, to evidence territoriality, and to consume the fruit of native figs, Ficus petiolaris (Mautz, 976, 982; Mautz and López- Forment, 978). In Oaxaca, specimens have been found in logs and rock crevices on Cerro Quiengola (Hartweg and Oliver, 94), and in caves and road culverts on the Isthmus of Tehuantepec (R.L. Bezy and colleagues, personal observations); in Chiapas, in logs and tree stumps in savanna (R.L. Bezy and colleagues, personal observations); and in Guatemala, on the rain forest floor and in rotting logs where the stomachs of specimens contained termites (Slevin, 942, as L. flavomaculatum). REMARKS. Walker (955) concluded that L. smithii is conspecific with L. flavimaculatum and this arrangement has been followed by several subsequent workers (e.g., Smith, 973). The ranges of the two forms are separated by only ca. 2 km on the Isthmus of Tehuantepec, with L. smithii extending north to Santo Domingo de Guzmán ( Santo Domingo Petapa; Goodwin, 969) and L. flavimaculatum south to El Mogoñé. In this region, the two species differ in number of femoral pores, lateral tubercle rows, and the presence of a pale interparietal spot. Although geographical variation among the localities of these 2 widely distributed species produces overlap in number of femoral pores, the two species consistently differ in two other scale features (Table ) and in chromosome number (Bezy, 972). The population of L. smithii found near Puerto Marqués, Guerrero, is the most divergent in scalation and its systematic status warrants further study. ETYMOLOGY. The name smithii is a noun in the genitive singular case and honors Sir Andrew Smith, M.D. ( ; founder of the South African Museum), whose description and figure of Poriodogaster grayi (a junior synonym of L. flavi-

32 3 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma Figure 28 Living individual of Lepidophyma sylvaticum (LACM 6744; snout vent length, 75 mm). maculatum) was used by Bocourt (876) to compare with L. smithii. COMMON NAMES. Smith s Tropical Night Lizard; Lagartija Nocturna de Smith. Lepidophyma sylvaticum Taylor Figure 28 Lepidophyma sylvatica Taylor, 939:, Figs., 2. Holotype: FMNH 2 (EHT-HMS 6259); type locality, 7 mi. north of Zacualtipán, Hidalgo. Gaigeia sylvatica: Smith, 942:38. Lepidophyma sylvaticum: Walker, 955:9. Lepidophyma flavimaculatum tenebrarum Walker, 955:. Holotype: UMMZ 374; type locality, 5 miles NW (by road) of Gómez Farías, in the Sierra Madre Oriental at Rancho del Cielo, 36 feet. DIAGNOSTIC CHARACTERS. The species differs from L. dontomasi, L. radula, L. tarascae, and L. lineri in having 22 or more femoral pores; from L. gaigeae in having distinctly enlarged lateral tubercles forming 5 to 38 rows; from L. tuxtlae and L. pajapanense in having the paravertebral rows composed of tubercles that are heterogeneous in size; from L. lowei in having 4 or more gulars; from L. occulor in having 56 or fewer gulars; from L. smithii in having a pale parietal spot throughout life; from L. lipetzi and L. chicoasense in having four or fewer pretympanics; from L. micropholis in having 27 or fewer dorsals; from L. reticulatum in having a higher second postorbital supralabial (RSLH of.84 or greater); and from L. flavimaculatum in having 3 or fewer pretympanics (99.3%) and a higher second supralabial (RSLH of.84 or greater; 99.3%; Table ). KARYOTYPE. Lepidophyma sylvaticum has a diploid chromosome number of 36 with 6 macrochromosomes and 2 microchromosomes. A microchromosomal heteromorphism was found in some individuals from Sierra de Cucharas (Bezy, 984). SIZE. The maximum snout vent length for the specimens examined is 3 mm. SEX RATIO. Of the 86 specimens sexed, 24 (28%) are males. Males appear to be rare in the populations found in the Sierra de Cucharas, Tamaulipas (3/29; %), and at El Lobo, Querétaro (2/3; 5%). DISTRIBUTION, HABITAT, AND NATURAL HISTORY. The species is found along the Sierra Madre Oriental from Nuevo León to Veracruz (Fig. 2). Outlying populations occur in canyons below the Cumbre de Monterey (Nuevo León), in the Sierra de Tamaulipas (Tamaulipas), the Sierra Álvarez (San Luis Potosí), and near Tlapacoyan and Al-

33 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma 3 varado, Veracruz. It has recently been reported from Puebla, near Tepango de Rodríguez (Camarillo, 995), and in the Sierra Norte (Canseco-Marquéz et al., 2). The species occurs primarily in pine oak woodland, humid pine forest, and cloud forest where it is found under bark, logs, and limestone rocks; and in rock crevices, walls, and buildings (Canseco-Marquéz et al., 2; Dixon et al., 972; Martin, 958; Walker, 955; R.L. Bezy and J.L. Camarillo, personal observations). It is sympatric with L. gaigeae in Querétaro. REMARKS. The occurrence of L. sylvaticum at Alvarado on the coast of Veracruz is somewhat enigmatic and brings the range of the species to within ca. 85 km of that of L. flavimaculatum. The two species appear to maintain their morphological differences in this area, with the Alvarado specimen (MZFC 722) differing from the northernmost L. flavimaculatum (ENEPI 24 from the Tuxtlas region) in pretympanics (2 vs 4) and the height of the second postorbital supralabial (RSLH.98 vs..58). Within L. sylvaticum, the morphologically most distinctive group of populations is composed of the samples from widely disjunct localities in the Sierra Álvarez of San Luis Potosí and in the northern base of the Sierra Madre Oriental of Nuevo León. ETYMOLOGY. The name sylvaticum (Latin: of woods) is an adjective and presumably refers to the occurrence of the species in humid pine and fir forest. COMMON NAMES. Madrean Tropical Night Lizard; Lagartija Nocturna de Montaña. Lepidophyma tarascae Bezy, Webb, and Álvarez Group S; Figure 29 Lepidophyma tarascae Bezy, Webb, and Álvarez, 982:, Figs. 3. Holotype: IPN 922; type locality, near Mexiquillo, Aquila District, Michoacán, México. DIAGNOSTIC CHARACTERS. The species differs from all other Lepidophyma except L. smithii and L. lineri in lacking a pale parietal spot in adults; from L. smithii, L. lineri, and L. occulor in having 59 or fewer dorsals and 43 or fewer gulars; from all except L. dontomasi, L. smithii, L. lineri, L. tuxtlae, and L. occulor in having 8 or fewer total femoral pores; and from all except L. radula, L. smithii, L. lineri, L. flavimaculatum, L. occulor, and L. sylvaticum in having 25 or fewer lateral tubercle rows (Table ). KARYOTYPE. Unknown. SIZE. The maximum snout vent length for the 4 specimens examined is 93 mm. SEX RATIO. Of the 4 specimens sexed, 3 (75%) are males. DISTRIBUTION, HABITAT, AND NATURAL HISTORY. The species is known only from two localities in Michoacán (Fig. 2): the type locality on the Pacific coast and the Cañón del Río Chilitos in the northern sector of the Sierra Coalcomán (Guzmán-Villa et al., 998). The type series was reported to have been collected in a rocky ravine in tropical semideciduous forest (Bezy et al., 982), but Álvarez and Díaz-Pardo (983) indicated that the specimens were found in encinar (evergreen oak woodland). The specimen from the Río Chilitos was taken from a cave in a transition zone between tropical deciduous forest and oak woodland (Guzmán-Villa et al., 998). REMARKS. Lepidophyma tarascae is most similar to L. smithii, and additional material may bridge the differences between the two in dorsals and gulars. All known specimens of L. tarascae have an uninterrupted row of enlarged tubercles extending from the posteroventral margin of the ear opening to above the gular fold (fig. 3 in Bezy et al., 982). However, in the Coalcomán specimen (MZFC 863) the row is not as distinct as in the type series (i.e., it is less clearly separated from the scattered large tubercles of the upper nuchal area) and approaches the condition seen in some L. smithii. ETYMOLOGY. The name tarascae is a noun in the genitive singular case and refers to the Tarasca, the indigenous people of the region of the type locality (Michoacán). COMMON NAMES. Tarascan Tropical Night Lizard; Lagartija Nocturna de Tarasca. Lepidophyma tuxtlae Werler and Shannon Group T3; Figure 3 Lepidophyma tuxtlae Werler and Shannon, 957: 9, Plate I. Holotype: UIMNH 6764 (FAS 55); type locality, lower slopes of Volcán San Martín, Veracruz, elevation 2,5 feet. Lepidophyma flavimaculatum tuxtlae: Savage, 963:35. Lepidophyma sawini Smith, 973:2, Plates 2. Holotype: UCM 4928; type locality, Vista Hermosa, Comaltepec, Ixtlán, Oaxaca, México. Lepidophyma alvarezi Smith, 973:5, Plates 2. Holotype: UCM 4928; type locality, 43 km on the road between Ocozocoautla and Mal Paso, 62 m Chiapas. DIAGNOSTIC CHARACTERS. The species differs from all other Lepidophyma except L. pajapanense in having the paravertebral rows composed of a nearly uninterrupted string of subequal tubercles; from L. pajapanense, L. lipetzi, and L. chicoasense in having 29 or fewer femoral pores; from all except L. mayae, L. pajapanense, L. lipetzi, and L. sylvaticum in having 3 or more large paravertebrals; and from L. radula, L. smithii, L. tarascae, L. lineri, and L. occulor in having 3 or more lateral tubercle rows (Table ). KARYOTYPE. Lepidophyma tuxtlae has a diploid chromosome number of 38 with 8 macrochromosomes and 2 microchromosomes (Bezy, 972). The karyotype appears to be identical to that of L. pajapanense, L. flavimaculatum, and L. reticulatum.

34 32 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma Figure 29 The holotype of Lepidophyma tarascae (IPN 922; snout vent length, 93 mm). SIZE. The maximum snout vent length for the 97 specimens examined is 97 mm. SEX RATIO. Of the 82 specimens sexed, 34 (4%) are males. Castillo-Cerón and López-González (99) found only one male (9%) in the specimens collected at their study site on Volcán Santa Marta, suggesting that the species may have a highly skewed sex ratio at certain localities. DISTRIBUTION, HABITAT, AND NATURAL HISTORY. The species is known from the Sierra de los Tuxtlas (Zapoapan to Volcán San Martín Tuxtla), Veracruz; the Sierra de Juárez, Oaxaca; the Río Coatzacoalcos and Sierra Atravesada on Isthmus of Tehuantepec, Veracruz and Oaxaca; and the Selva del Ocote, Chiapas (Fig. ). It is sympatric with L. flavimaculatum in the Río Coatzacoalcos and the Ocote areas and with L. pajapanense in the Tuxtlas and Río Coatzacoalcos regions. Vogt et al. (997) indicated that in the Tuxtlas the species is nocturnal and inhabits trees within the rain forest from sea level to 5 m. Individuals have been found within rotting logs and beneath the bark of tree stumps within or near the rain forest (R.L. Bezy and C.J. Cole, personal observations). Greene (97) reported that a female collected near Lago Catemaco in June contained 6

35 Contributions in Science, Number 493 Bezy and Camarillo: Systematics of Lepidophyma 33 Figure 3 Living individuals of Lepidophyma tuxtlae from the Tuxtlas region of Veracruz (upper; LACM 36358; snout vent length, 62 mm) and the Vista Hermosa region of Oaxaca (lower; UMMZ 2587; snout vent length, 97 mm; photo by Kraig Adler).